The characteristics of the fish and assemblages of the Leschenault . Inter-period comparisons and their management implications.

Lauren Veale, Peter Coulson, Steeg Hoeksema, James Tweedley,

Norm Hall and Ian Potter

Centre for Fish and Research, Murdoch University

South West Development Commission Report

December 2010

The characteristics of the fish and crab assemblages of the

Leschenault Estuary. Inter-period comparisons and their

management implications.

Lauren Veale, Peter Coulson, Steeg Hoeksema, James Tweedley, Norm Hall and Ian Potter*

*Contact author Email: [email protected] Centre for Fish and Fisheries Research School of Biological Sciences and Biotechnology Murdoch University South Street, Murdoch Western , 6150 Ph: 9239 8801 Fax: 9360 6303

This work is copyright. Except as permitted under the Copyright Act 1968 (Cth), no part of this publication may be reproduced by any process, electronic or otherwise, without the specific written permission of the copyright owners. Neither may information be stored electronically in any form whatsoever without such permission.

2

Table of Contents Executive Summary ...... 4 Introduction ...... 10 Materials and methods ...... 15 Sampling of fishes and ...... 15

Treatment of samples ...... 17

Environmental variables ...... 17

Assigning life cycle guilds to fishes ...... 18

Statistical analyses – fish community ...... 18

Statistical analyses – Blue Swimmer Crab assemblage ...... 22

Results ...... 23 Environmental variables ...... 23

Abundances and numbers of fish ...... 23

Densities of commercial and recreational fish species...... 27

Numbers and percentage contributions of the different life cycle guilds ...... 27

Compositions of the fish faunas ...... 29

Numbers of species and densities of fish ...... 39

Sex ratio and minimum legal carapace width for Blue Swimmer Crabs ...... 42

Densities and catch rates of Blue Swimmer Crabs ...... 42

Carapace width frequency distributions for Blue Swimmer Crabs ...... 43

Historical comparisons of densities and catch rates of Blue Swimmer Crabs ...... 48

References ...... 54

3 Executive Summary

A total of 14,200 fishes was caught in the lower and middle regions of the Leschenault

Estuary using a 21.5m seine net in each season between winter 2008 and autumn 2010. This total was only 3% less than the 14,601 fishes caught using the same seine net at the same sites twice seasonally in 1994, i.e. with the same amount of fishing effort. The numbers of species recorded in 2008-10 (36) and 1994 (33) were also similar. The above absence of a marked difference in the abundance of fish is consistent with the similarity in the mean densities of fishes per sample in the two periods. However, the mean number of species per sample and measures of diversity were greater in the current than earlier period.

The eight most abundant species in 2008-10 ranked among the 11 most abundant species in 1994. Furthermore, the five most abundant species in 1994, which collectively accounted for

~ 90% of the total number of fish caught in that period, ranked amongst the top six species in

2008-10, recognising, however, that they contributed less, i.e. ~ 69%, to the total number of fish in that later period. These five species were the Elongate Hardyhead Atherinosoma elongata, the

Sandy Sprat Hyperlophus vittatus, the Yelloweye Mullet Aldrichetta forsteri, the Silverfish

Leptatherina presbyteroides and the Southern Longfin Goby Favonigobius lateralis.

The far greater contributions of the most abundant species in 1994 than 2008-10 reflects the extreme dominance of the Southern Longfin Goby and Sandy Sprat in the earlier period.

These two species thus contributed nearly 70% to the total number of fishes caught in the earlier period, compared with only 35% by the two most abundant species, i.e. Elongate Hardyhead and

Sandy Sprat, in the later period. This helps account for the diversity of the fish fauna being less in the earlier period.

4 In contrast to the very high ranking of the Elongate Hardyhead, the Sandy Sprat, the

Yelloweye Mullet, the Silverfish and the Southern Longfin Goby in both periods, the Spotted

Hardyhead Craterocephalus mugiloides and the Common Hardyhead Atherinomorous vaigiensis ranked fifth and ninth, respectively, in terms of abundance in 2008-10, but were not caught in

1994. Furthermore, these two species of Hardyhead were likewise relatively abundant in additional samples taken in the Leschenault Estuary using a 41.5 m seine in 2008-10 and were never caught using that net in 1994. As these two Hardyheads are tropical species, and have not previously been recorded south of the Peel-Harvey Estuary, their current substantial numbers in the Leschenault Estuary represent a southwards extension of their distribution, presumably in response to the increases in coastal water temperatures that have occurred along the lower west during recent years (Pearce and Fang, 2007).

The above inter-period differences help account for the species compositions in 1994 and

2008-10 being significantly different. However, the most important species for distinguishing between the two periods was the Southern Longfin Goby, which ranked first by abundance and contributed 36.5% to the total number of fishes caught in 1994, compared with ranking only sixth and contributing only 8.1% in 2008-10. Furthermore, the densities of the other two abundant species of goby, the Bluespot Goby Pseudogobius olorum and particularly the

Southwestern Goby Afurcagobius suppositus, were also greater in the earlier period. The fact that these species live in and/or on the substrate suggests that the benthic environment may have deteriorated during recent years. Alternatively, the conditions in 1993 and 1994 may have been particularly favourable for the successful spawning and recruitment of goby species.

Among the commercial and recreational species, the overall densities of the King George

Whiting Sillaginodes punctata, the Yellow-fin Whiting schomburgkii and the Sea Mullet

5 Mugil cephalus did not differ significantly between periods, whereas that of the Yelloweye

Mullet was greater in 2008-10 than in 1994. The Western Australian Salmon Arripis truttaceus was moderately abundant in 2008-10 but was not caught in the earlier period.

A total of 500 Blue Swimmer Crabs, Portunus pelagicus, was caught in the Leschenault

Estuary using 21.5 and 41.5 m long seine nets over two years and a further 194 were obtained using pots in the last of those years. The carapace width of only 59 of those crabs (9.4%) was greater than the minimum legal carapace width for retention (127 mm) of this species. All but five of the crabs with a carapace width greater than 127 mm were males, reflecting the tendency for the males of this marine species to remain longer in the estuary. The Blue Swimmer Crabs exhibited considerable variation in the times and strengths of recruitment. Thus, for example, a large number of crabs entered the estuary in the autumn of 2010 but not in that season in 2009.

Furthermore, the catches taken from seine nets and crab pots both showed that recruitment of crabs into the estuary was greater in 1996/97 and 2009/10 than in 1997/98. Such interannual variations in estuarine recruitment are just as likely to be due to the influence of conditions in marine waters, where the eggs develop into zoeae and then juveniles, as they are to conditions in the estuary itself.

Main findings and conclusions

1) The following inferences regarding similarities and differences in the fish faunas of the

Leschenault Estuary in two different periods are based on comparisons between data for two years in the more recent period (2008/09, 2009/10) and those for only a single year in the earlier period (1994). As most of the dominant species found in have short life cycles and the extent of their recruitment can thus vary markedly among years and thereby lead to differences

6 in species composition, even between successive years, caution should be exercised when drawing inferences from our inter-period comparisons.

2) Despite the above caveat, there were some clear and valid similarities between the characteristics of the fish faunas in 1994 and 2008-10. For example, both the total number of individuals and total number of species caught in the two periods using the same seine net and effort were very similar and eight of the most abundant species in 2008-10 ranked among the 11 most abundant species in 1994.

3) The compositions of the fish fauna of the Leschenault Estuary in 1994 and 2008-10 did differ however. This difference was largely attributable to the colonisation and success of two species of Hardyhead in the intervening period and to the decline in the abundance of three species of gobies. The colonisation by the Spotted and Common Hardyheads represents a southwards extension of these tropical species, presumably reflecting the influence of increasing coastal water temperatures during recent years, whereas the declines in the abundances of the Southern

Longfin Goby, Bluespot Goby and Southwestern Goby suggests that the quality of the benthic environment may have deteriorated. Alternatively, in the case of the goby species, conditions in the benthos in the early 1990s may have been particularly conducive to the successful spawning and recruitment of these benthic species.

4) The overall mean diversity and evenness of the fish fauna were greater in 2008-10 than in

1994, reflecting, in part, the colonisation of the estuary by substantial numbers of two species of

Hardyhead and a reduction in the extreme dominance of the most abundant goby species.

7 5) The densities of commercial and recreational fish species (Yelloweye Mullet, Sea Mullet,

Yellowfin Whiting and ) in nearshore waters were low and did not exhibit an overall tendency to increase or decrease between 1994 and 2008-10.

6) On the basis of data derived from seine netting in the two seasons when crab abundance was greatest (summer and autumn), the densities of Blue Swimmer Crabs in 1996/97 and 2009/10 were appreciably greater than in 1997/98 and 2008/09. Furthermore, the catch rates derived from catches in pots were also greater in 1996/97 and 2009/10 than in 1997/98. These data demonstrate that the strength of recruitment into the estuary varies among years and that the abundance in the most recent year of sampling was relatively high.

7) In summary, the only conclusive change to the fish fauna of the Leschenault Estuary is the colonisation of the estuary in recent years by two tropical species of Hardyhead. The indications that the densities of three goby species may have declined, possibly as a consequence of the detrimental effects of changes to the benthos, contrasts with the suggestion from the increase in diversity that the estuarine environment may even have improved between the two periods.

8) The species composition of the fish fauna changed between 1994 and 2008-10 in response to climatic and presumably also anthropogenic factors. There is no indication, however, that the current risk status of moderate, as identified by the Department of Fisheries, requires alteration.

9) There is a need to monitor, in a consistent manner, the fish fauna and crab assemblages at regular intervals in the future so that any changes in their characteristics can be detected and thus enable the reasons for those changes, and whether they are due to anthropogenic and/or climatic

8 effects, to be elucidated. This will enable managers to put in place any necessary initiatives to counteract the effects of those changes.

10) In view of the possibility that the benthic environment of the Leschenault Estuary may be deteriorating, it is very strongly recommended that a rigorous study of the benthic macroinvertebrates is undertaken to ascertain whether the characteristics of this fauna indicate that this is the case. Comparisons between the benthic macroinvertebrate faunas in the Peel-

Harvey and Swan-Canning estuaries in the 1980s and 2000s provided very strong evidence that, particularly in the former estuary, the condition of the benthic environment declined between those periods.

11) From a management point of view, the changes in the species composition of the fish fauna and the abundances of the various fish species and crabs within the Leschenault Estuary are not of a nature, consistency or magnitude that would indicate definitively that the quality of the estuarine environment for the fish and crab faunas has changed markedly over the last 15 years.

While there is thus no strong indication that current management regulations require modification, the possibility that the benthic environment has deteriorated (see point 10) requires investigation.

9 Introduction

The Leschenault Estuary, which is located at 115° 42‟E, 33° 16‟S approximately 150 km south of the city of Perth, is a 14 km long shallow, lagoonal-like water body that lies parallel to the coast (Semeniuk and Meagher, 1981; Fig. 1). The Leschenault Estuary has a surface area of

25 km2, an average depth of < 1m and a maximum depth of 2 m (Semeniuk and Meagher, 1981;

Brearley, 2005). The estuary receives freshwater input from the Collie and Preston rivers at the southern end of the basin, which are opposite the permanently-open and artificial entrance channel, and from the Parkfield Drain at the northern end of the basin (Brearley, 2005;

McKenna, 2007). The catchment of the Leschenault Estuary, comprising an area of ~ 2,020 km2, includes catchments for the Wellesley, Brunswick, Ferguson rivers and the Collie River catchment below Wellington Dam (Kelsey and Hall, 2010). Although a large area in the upper reaches of this catchment has retained its natural vegetation, the Swan Coastal Plain and land to the east of the Darling Scarp support a diverse range of agricultural and other land uses, including dairy and beef cattle farming, horticulture, forestry and mining, manufacturing and service industries, and is also the centre of a rapidly expanding urban region (McKenna, 2007;

Kelsey and Hall, 2010). The intense land use has lead to the eutrophication of the catchment waterways, and in particular of the lower Collie and Brunswick rivers, where an increased input of nutrients has led to the periodic development of algal blooms and fish kills (Kelsey and Hall,

2010).

The hydrology of the Leschenault catchment and estuary became greatly modified when, in 1933, the Wellington Dam on the Collie River was constructed and, in 1951, the natural entrance channel at the bottom end of the estuary was closed and a new and permanent opening was developed opposite the mouth of the Collie River and this opened into the northern limit of

10 Koombana (McKenna, 2000, 2007; Brearley, 2005). Subsequent in-filling led to the isolation of the lower area of the estuary basin from the main body of that system. The absence of riverine discharge into the middle and upper regions of the basin, allied with the shallowness of those regions, led to the waters in those regions becoming hypersaline in late summer and autumn in most years (Potter and de Lestang, 2000). The modification of tributaries, i.e. de- snagging and straightening, and the construction of drainage channels have greatly modified the catchment‟s natural drainage, which is dominated by river flow from winter rains and scheduled water releases from Wellington Dam during those months (Beckwith 2008; Kelsey and Hall,

2010).

Recent results of research carried out by the Department of Water have shown that the substrate of the Leschenault Estuary has a greater proportion of fine sediment (i.e. mud) than that of the Peel-Harvey and Swan-Canning estuaries, and that this decreases with increasing distance from the mouths of the Collie and Preston rivers (Kilminster, 2010). The transport of sediment from cleared catchment areas and the wave-dominated nature of the Leschenault Estuary promote sediment retention, with the central basin acting as a sink for fine sediments (McKenna,

2007). Increased sedimentation in the Leschenault Estuary could have an impact on the aquatic fauna, such as benthic fishes, by “smothering” their habitat, clogging their gills and reducing their feeding efficiency and food quality (Gill and Potter, 1993; McKenna, 2007). Concentrations of nutrients, namely phosphorus and nitrogen, recorded in the sediment in the late 1980s and

2008 are very similar (cf. McComb et al., 2000; Kilminster, 2010)

In 1999, the Leschenault Estuary was the focus of a symposium aimed at drawing together data obtained from numerous studies on the floras, faunas and physico-chemical characteristics of the Leschenault Estuary and discussing their significance and implications for

11 understanding the functioning of this system (e.g. Hillman et al., 2000; Potter and de Lestang,

2000; Potter et al., 2000; McComb et al., 2000; Raines et al., 2000). The results of these studies and their significance were reported in a special edition of the Journal of the Royal Society of

Western Australia that was published in 2000. This special edition included the results of studies on the composition of the fish fauna of the Leschenault Estuary (Potter et al., 2000) and on the biology of the Blue Swimmer Crab Portunus pelagicus in the estuary and adjacent marine waters

(Potter and de Lestang, 2000).

In summary, the study by Potter et al. (2000) on the fish assemblages of the nearshore waters of the lower and middle regions of the Leschenault Estuary yielded 42 fish species. The four most abundant species were the Southern Longfin Goby Favonigobius lateralis, the Sandy

Sprat Hyperlophus vittatus, the Silver Fish Leptatherina presbyteroides and the Elongate

Hardyhead Atherinosoma elongata, which collectively contributed 83% to the total number of fish caught (Potter et al., 2000). The catches, in this earlier period, of commercially/recreationally important species such as Yelloweye Mullet Aldrichetta forsteri,

Sea Mullet Mugil cephalus, King George Whiting Sillaginodes Punctata and Yellow-fin Whiting

Sillago schomburgkii, contributed < 6% of the overall fish catch in the estuary (Potter et al.,

2000).

As with the Swan-Canning and Peel-Harvey estuaries just to the north on the lower west coast of Australia, the Leschenault Estuary acts as a nursery area for a number of marine species

(Potter and Hyndes, 1999; Potter et al., 2000). Indeed, of the 42 fish species caught in 1994, 20 are marine species that use the estuary as a nursery area, i.e. marine-estuarine opportunists

(Potter et al., 2000). Although numerous in terms of the number of species, these marine species contributed only 32% to the total number of fish caught. In comparison, the 13 species able to

12 complete their life cycles in the Leschenault Estuary, i.e. the estuarine species, were far more abundant, contributing 68% to the total number of fishes in this system.

Potter and de Lestang (2000) showed that, in the Leschenault Estuary, the abundance of

Blue Swimmer Crabs increases in the summer and autumn months and declines markedly in the subsequent winter and spring months when water temperatures and salinity decline precipitously.

In the case of both the fish fauna and the crab population, these previous studies were carried when there was an active commercial for fishes and Blue Swimmer Crabs in the

Leschenault Estuary. However, since the closure of these commercial fisheries in 2001, no research has been conducted on either the fish communities or the assemblages of Blue Swimmer

Crab in the estuary.

The agreed outcomes of the current study were as follows.

Production of a report on the current densities of crabs and fishes in the Leschenault

Estuary showing the diversity and species composition of the fish fauna. A comparison of these results with those recorded for Leschenault Estuary in the mid 1990s to determine the current status of the fauna relative to that observed in the earlier period and advise the potential implications for the local community and resource managers of any changes in the fauna that are detected.

The aims of this project were as follows

1) Determine the number of overall species, densities and composition of the fish fauna in nearshore, shallow waters of the Leschenault Estuary in each season between winter 2008 and autumn 2010.

13 2) Compare the data for species richness, densities and compositions of the fish fauna in different regions and seasons in 2008-10 with those obtained twice seasonally in 1994, i.e. with the same sampling effort. These comparisons will elucidate whether there are any clear indications that the characteristics of the fish fauna in the Leschenault Estuary have changed markedly during the past 15 to 20 years.

3) Compare the densities of recreationally and commercially important fish species recorded in nearshore, shallow waters of this system in 2008-10 with those in the early 1990s.

4) Determine the abundances and distributions of Blue Swimmer Crabs, Portunus pelagicus, in nearshore, shallow and offshore, deeper waters of the estuary.

5) Compare the above data for Blue Swimmer Crabs in 2008-10 with those obtained seasonally in 1996-98. These comparisons will be used to elucidate whether there are any clear indications that the abundances of Blue Swimmer crabs in the Leschenault Estuary have changed between those periods.

6) Discuss the significance of inter-period comparisons for the management of the Leschenault

Estuary and of any changes resulting from climate change or anthropogenic effects.

14 Materials and methods

Sampling of fishes and crabs

The fishes in nearshore, shallow waters of the lower and middle regions of the

Leschenault Estuary were sampled by seine net in each season over two consecutive years (12 months) from July (winter) 2008 to April (autumn) 2010. These samples were collected from four sites in both the middle and lower regions of this estuary, with three of these sites in each region being the same as those sampled in 1994 (Fig. 1; Potter et al., 2000). The 21.5 m long seine net, which had the same dimensions as that used in the earlier study (Potter et al., 2000), comprised two 10 m long wings (6 m of 9 mm mesh and 4 m of 3 mm mesh) and a 1.5 m bunt (3 mm mesh). The net swept an area of 116 m2 and fished to a maximum depth of 1.5 m. Blue

Swimmer Crabs, Portunus pelagicus, in nearshore waters were also sampled using a 41.5 m seine net, which consisted of two 20 m long wings (25.4 mm mesh) and a 1.5 m bunt (9 mm mesh) and swept an area of 274 m2. Blue Swimmer Crabs were also sampled in the upper and apex regions of the estuary (Fig. 1), which were not sampled for fishes in the earlier study.

Blue Swimmer Crabs in offshore, deeper waters of the lower and middle regions of the estuary were sampled seasonally using crab pots baited with the Australian Herring Arripis georgianus. As in the earlier study of crabs in the Leschenault Estuary (Potter and de Lestang,

2000), the pots used in the present study were 630 mm high, 1000 mm in diameter and made of

76 mm mesh. On each sampling occasion at each site, four crab pots were set in a row at approximately 15 m intervals for 24 h.

15

Figure 1. Map showing sites in the lower and middle regions of the Leschenault Estuary that were sampled for fish and crabs using a 21.5 m seine net nearshore waters (circles) and crab pots in offshore waters (squares). Sites in nearshore waters of the upper and apex regions were also sampled for crabs using a 21.5 m seine net. Black circles represent sites sampled for fish in 1994 and 2008-10, while yellow circles represent additional seine netting sites sampled for fish in 2008-10. Black squares represent sites sampled for Blue Swimmer Crabs in 1996-98 and 2008- 10. Insets show location of Leschenault Estuary in south-.

16 Treatment of samples

Following their capture, fishes were immediately placed in an ice slurry and transported to the laboratory, where they were identified to species and the total number of individuals and biomass (to the nearest 0.1 g) of each species in each sample were recorded. The total length

(TL) of each fish was measured to the nearest 1 mm, except when the total number of individuals of a species in a sample was large, in which case the lengths of a random subsample of 50 individuals were measured. The total number of species in each replicate sample was converted to a density, i.e. number of individuals 100 m-2.

All of the Blue Swimmer Crabs collected from both nearshore, shallow waters using the seine net and from deeper, offshore waters using crab pots were measured, sexed and immediately returned to the water. The carapace width (CW) of each crab, i.e. the distance between the tips of the two lateral spines of its carapace, was recorded to the nearest 1 mm using vernier callipers. The total number of crabs collected in each replicate sample from inshore waters using the seine net was converted to a density (crabs 100 m-2), while that collected in each replicate sample from deeper waters using crab pots is expressed as number of crabs pot-1day-1.

Environmental variables

Salinity and water temperature were recorded at each site on each sampling occasion using a Yellow Springs International Model 85 oxygen, conductivity, salinity and temperature meter. These environmental variables were measured in the middle of the water column at each nearshore site sampled by seine net and at the top and bottom of the water column at each offshore site where crab pots were deployed.

17 Assigning life cycle guilds to fishes

Each species was allocated a life-cycle guild (Elliott et al., 2007), based on the results of numerous studies on the biology of the various fish species found in south-western Australian estuaries (Potter and Hyndes, 1999). Those species that at sea, but typically enter estuaries in large numbers and particularly as juveniles, are termed marine-estuarine opportunists (MO), while those that spawn at sea and typically enter estuaries only in low and often irregular numbers are referred to as marine stragglers (MS). Species that are only found in estuaries are termed estuarine (E), while those that complete their life cycle in estuaries but are also represented by discrete marine populations are termed estuarine and marine (E&M). The percentage contributions of each of these life-cycle guilds, in terms of both the number of species and number of individuals, were determined for both the lower and middle regions of the estuary.

Statistical analyses – fish community

The statistical analyses used in this study employed the routines in the PRIMER v6 multivariate statistics package (Clarke and Gorley, 2006) and the PERMANOVA+ add-on module (Anderson et al., 2008).

Initially, the species compositions of the fish assemblages in the lower and middle regions of the Leschenault Estuary in the summer, autumn, winter and spring of 1994 were compared with those in four consecutive seasons between winter 2008 and autumn 2009 and between winter 2009 and autumn 2010. For convenience, each of the four consecutive seasons are considered to represent a year and are thus subsequently referred to as 1994, 2008/09 and

2009/10. Note that, as sampling was undertaken twice in each season in 1994 and once in both

18 2008/09 and 2009/10, the data for 1994 were restricted to those collected on one of the two sampling occasions to enable the comparisons between the earlier and later periods to be based on the same number of samples.

The abundances of each fish species at the corresponding three sites (replicates) in both the lower and middle regions in each season in each year were initially subjected to dispersion weighting to down-weight the contributions of those species that exhibited erratic differences in abundance between replicate samples (Clarke et al., 2006). The resultant data were then square- root transformed to reduce the influence of any abnormally high abundances for one or more species. These transformed data were then used to construct a Bray-Curtis similarity matrix, which was subjected to PERMANOVA. These data are considered to represent a 3-way crossed design, comprising year (3 levels, i.e. 1994, 2008/09 and 2009/10) × season (4 levels, i.e. each season of the year) × region (2 levels, i.e. lower and middle estuary), with each of these factors being fixed. This PERMANOVA test was employed to investigate whether there were interactions between the above three factors and, if so, to determine the extent of those interactions relative to each other and to those of the main effects.

As PERMANOVA detected significant interactions between the ichthyofaunal compositions in each year with both season and region (see Results), two-way crossed Analysis of Similarities (ANOSIM; Clarke and Green, 1988) tests, employing the above matrix, were conducted to examine the influence of each factor in the year × season and year × region interactions. ANOSIM was preferred at this stage of the analysis because of the greater robustness of a fully non-parametric test and the fact that the variations among the R-statistic values provide a measure of the relative differences between the factors. In this and all subsequent ANOSIM tests, both the R-statistic values and their associated p values are given,

19 with small values of p, typically < 5%, being interpreted as casting doubt on the null hypothesis that the fish compositions were not significantly different. The magnitude of the R-statistic typically ranges from 1, when all samples within each group are more similar to each other than to any of the samples from other groups, down to ~ 0, when the average similarity among and within groups do not differ (Clarke and Gorley, 2006).

Non-metric Multidimensional Scaling (nMDS; Clarke et al., 1993) ordination plots were constructed from Bray-Curtis Similarity matrices, derived from the mean densities of the various fish species in the lower and middle regions of the estuary in each successive season in 1994,

2008/09 and 2009/10. These plots were used to visualise the extent to which the a priori groups of a factor differ and the basis for any interactions between factors, as identified by

PERMANOVA.

As the results of the above analyses demonstrated that the fish compositions in 2008/09 and 2009/10 were not significantly different (see Results), the abundances of fish in each region in each season were pooled for the two years. The pooled years are subsequently referred to as

2008-10. The abundances of fishes recorded in the first and last months of each season in 1994 in each region were also pooled. Thus, the comparisons between the earlier and later periods are based on the same number of samples, i.e. 24. The abundances of the various fish species in each region in each season of each period were used to derive a Bray-Curtis Similarity matrix, which was then subjected to a 3-way crossed PERMANOVA design comprising period (2 levels, i.e. 1994 vs 2008-10) × season (4 levels, i.e. each season of the year) × region (2 levels, i.e. lower and middle regions of the estuary), with each factor being considered fixed. As PERMANOVA detected significant and strong interactions between period and season and between season and

20 region (see Results), two-way crossed ANOSIM tests employing the above matrix were conducted to examine the influence of each factor in those interactions.

A two-way crossed Similarity Percentages analyses (SIMPER, Clarke and Gorley, 2006) was then used to identify which fish species typified the faunal composition of each period and season and, which were responsible for distinguishing between the fish compositions in each pair of those groups. In the case of the period × season interaction, the species driving seasonal differences were examined separately for each period employing a one-way SIMPER. The focus in all SIMPER analyses was placed on those species that had the highest similarity/standard deviation and dissimilarity/standard deviation ratio and were most abundant.

The number of species and density of fish at each of the replicate sites in the lower and middle regions of the Leschenault Estuary in each calendar season of the two periods, i.e. 1994 and 2008-10, were used to construct separate Euclidean distance matrices, which were each subjected to a single variable three-way PERMANOVA. The results were used to determine whether these two biotic variables were influenced by period, season and/or region, each of which was considered a fixed factor. Prior to subjecting the number of species and density of fish to PERMANOVA, they were square root and log10(n) transformed, respectively. These transformations were shown to be appropriate by examining the relationship between the log10 of the mean of the replicates in each region in both periods (see Clarke and Warwick, 2001 for rationale and details of this approach). When interactions were present in the case of either the mean number of species or densities of fishes, figures were plotted to enable the basis for those interactions to be visualised.

The overall mean number of fish species, density of individuals, Pielou‟s Evenness,

Simpson‟s Index and quantitative taxonomic distinctness were calculated for each sample in each

21 period. The last variable is a measure of species diversity that accounts for the taxonomic relatedness of individuals from different species in a sample, based on their level of taxonomic separation through the hierarchical levels of the Linnaean tree (Warwick and Clarke, 1995).

Statistical analyses – Blue Swimmer Crab assemblage

The densities of Blue Swimmer Crabs in nearshore, shallow waters, derived from catches obtained using a 21.5 m seine net in the lower and middle regions of the Leschenault Estuary in four consecutive seasons between winter and autumn in 1996/97 and 1997/98 were compared with those in the same seasons in 2008/09 and 2009/10. For convenience, each of the four consecutive seasons are considered to represent a year and are thus subsequently referred to as

1996/97, 1997/98, 2008/09 and 2009/10.

The catch rates of Blue Swimmer crabs in deeper, offshore waters, derived from catches obtained using pots in winter to autumn in 1996/97 and 1997/98, were compared with those obtained for the corresponding four seasons in 2009/10.

The densities and catch rates of Blue Swimmer Crabs at each site (replicate) in both the lower and middle regions of the Leschenault Estuary in each season and year were square-root transformed to reduce the influence of any abnormally high densities or catch rates and then used to construct separate Euclidean distance matrices, which were subjected to 3-way crossed

PERMANOVA. The PERMANOVA design, in the case of nearshore waters, comprised period

(4 levels, i.e. 1996/97 vs 1997/98 vs 2008/09 vs 2009/10) × season (4 levels, i.e. each season of the year) × region (2 levels, i.e. lower and middle regions of the estuary), with each factor being considered fixed. In the case of offshore waters, period was represented by 3 levels (i.e. 1996/97 vs 1997/98 vs 2009/10). This PERMANOVA test was employed to investigate whether there

22 were interactions between the above three factors and, if so, to determine the extent of those interactions relative to each other and to those of the main effects.

Results

Environmental variables

The mean seasonal salinities in the lower region of the estuary in 1994, 2008/09 and

2009/10 varied (Fig. 2a, b). Mean salinities in 1994 and 2008/09 rose from their minima in winter (~ 25, 24, respectively) to their maxima in summer (~ 34, 36, respectively), whereas, in autumn, they then rose slightly in 2008/09 and declined markedly in 1994. In contrast, mean seasonal salinities in the lower region during 2009/10 showed relatively little seasonal variation, ranging from ~ 29 in spring to ~ 32 in autumn (Fig. 2a, b). The mean salinities in the middle region followed a consistent pattern in all three years, with the seasonal values in each year rising to their maxima in summer and then declining in autumn. In three of the four seasons, the mean seasonal salinities were greatest in 2008/09, followed by 2009/10 and then 1994 (Fig. 2a, b).

The mean seasonal water temperatures in the lower and middle regions of the

Leschenault Estuary in 1994, 2008/09 and 2009/10 all rose from their minima of 16.5, 14.1, and

14.3 °C in winter to their maxima of 26.0, 27.5, and 27.8 °C in summer, respectively (Fig. 2c, d).

Abundances and numbers of fish species

A total of 12,241 fishes were recorded during seasonal sampling of the lower and middle regions of the estuary between winter 2008 and autumn 2010, after the number of fish in each

23 sample had been adjusted to a constant density, i.e. number of fish 100 m-2, and summed (Table

1). This is similar to the 12,587 fish recorded using samples collected twice seasonally between summer and spring 1994 (i.e. using the same sampling effort) and similarly corrected (Table 1).

The mean numbers of fish 100 m-2 in the current (255) and earlier periods (262) were similar.

Figure 2. Mean seasonal salinities (a, b) and water temperatures (c, d) ± 1 SE recorded in the nearshore waters of the lower and middle regions of the Leschenault Estuary at the time of sampling in 1994 (black circles), 2008/09 (grey circles) and 2009/10 (white circles).

24 Table 1. Comparisons between the number (N), ranking (R) and percentage contribution (%) of each fish species caught collectively in the lower and middle regions of the Leschenault Estuary using the 21.5 m seine net in 1994 and 2008-10. Number of each species (N) in both time periods and overall were calculated from the number of individuals in each sample, following their adjustment to a constant density, i.e. number of fish 100 m-2. The actual total number of all fish caught is also given. Life-cycle guilds, ER = estuarine resident, E&M = estuarine and marine, MO = marine estuarine-opportunist, MS = marine straggler.

Common name Species name Guild 2008/2010 1994 N R % N R % Elongate Hardyhead Atherinosoma elongata ER 2220 1 19.1 521 4 4.1 Sandy Sprat Hyperlophus vittatus MO 1868 2 16.0 4084 2 32.4 Yelloweye Mullet Aldrichetta forsteri MO 1478 3 12.7 212 5 1.7 Silver Fish Leptatherina presbyteroides E&M 1473 4 12.7 1985 3 15.8 Spotted Hardyhead Craterocephalus mugiloides ER 1218 5 10.5 Southern Longfin Goby Favonigobius lateralis E&M 947 6 8.1 4597 1 36.5 Western Gobbleguts Apogon rueppelli E&M 846 7 7.3 136 9 1.1 King George Whiting Sillaginodes punctata MO 261 8 2.2 191 7 1.5 Common Hardyhead Atherinomorus vaigiensis MO 260 9 2.2 Weeping Toadfish Torquigener pleurogramma MO 228 10 2.0 68 12 0.5 Western Hardyhead Leptatherina wallacei ER 222 11 1.9 138 8 1.1 Soldier Gymnapistes marmoratus MO 205 12 1.8 38 14 0.3 Western Australian Salmon Arripis truttaceus MO 78 13 0.7 Blue Weed-whiting Haletta semifasciata MO 73 14 0.6 11 19 0.1 Bluespot Goby Pseudogobius olorum ER 61 15 0.5 126 10 1.0 Sea Mullet Mugil cephalus MO 47 16 0.4 108 11 0.9 Elongate Flounder Ammotretis elongatus MS 39 17 0.3 2 27 <0.1 Spotted Pipefish Stigmatopora argus MS 21 18 0.2 40 13 0.3 Sillago schomburgkii MO 14 19 0.1 4 23 <0.1 Western Trumpeter Whiting Sillago burrus MO 13 20 0.1 19 16 0.2 Rough Leatherjacket Scobinichthys granulatus MS 13 20 0.1 3 26 <0.1 Tarwhine Rhabdosargus sarba MO 8 22 0.1 6 21 <0.1 Hairy Pipefish Urocampus carinirostris E&M 5 23 <0.1 17 17 0.1 Western Striped Grunter Pelates octolineatus MO 4 24 <0.1 5 22 <0.1 Smalltooth Flounder Pseudorhombus jenynsii MO 3 25 <0.1 4 23 <0.1 Southwestern Goby Afurcagobius suppositus ER 3 26 <0.1 197 6 1.6 Prickly Toadfish Contusus brevicaudus MO 3 26 <0.1 38 14 0.3 Wood's Siphonfish Siphamia cephalotes E&M 3 26 <0.1 16 18 0.1 Tailor Pomatomus saltatrix MO 2 29 <0.1 1 30 <0.1 Common Silverbiddy Gerres subfasciatus MO 2 29 <0.1 Sixspine Leatherjacket Meuschenia freycineti MO 2 29 <0.1 Blue Sprat Spratelloides robustus MO 1 32 <0.1 4 23 <0.1 Bridled Goby Arenigobius bifrenatus E&M 1 32 <0.1 1 30 <0.1 Rock Flathead Platycephalus laevigatus MS 1 32 <0.1 Flathead Goby Callogobius depressus E&M 1 32 <0.1 Serpent Eel Ophisurus serpens MO 1 32 <0.1 Old Wife Enoplosus armatus MS 10 20 0.1 Southern School Whiting Sillago bassensis MS 2 27 <0.1 Sculptured Goby Callogobius mucosus MO 2 27 <0.1 Brownspotted wrasse Notolabrus parilus MS 1 30 <0.1 Southern Crested Weedfish Cristiceps australis MS 1 30 <0.1 Table 1 continued on next page

25 Table 1 continued

2008/10 1994 Total number of samples 48 48 Total number of species 36 33 Total number of fish 14,200 14,601 Total number of fish after correction to a constant density 12,241 12,587 Mean density of fish 100 m-2 (± SE) 255.0 43.4 262.2 68.8 Mean number of species per sample (± SE) 7.3 0.475 4.9 0.420 Mean Taxonomic Distinctness per sample (± SE) 58.6 1.34 53.7 2.92 Mean Pielou’s Evenness per sample (± SE) 0.577 0.03 0.463 0.04 Mean Simpson’s Index per sample (± SE) 0.527 0.03 0.376 0.04

In 1994, the fish fauna was highly dominated by the Southern Longfin Goby

Favonigobius lateralis (36.5%) and the Sandy Sprat Hyperlophus vittatus (32.4%), which thus

collectively contributed nearly 70% to the total fish catch (Table 1). The third most abundant

species was the Silverfish Leptatherina presbyteroides, which contributed 15.8% to the total

catch, while the Elongate Hardyhead Atherinosoma elongata and Yelloweye Mullet Aldrichetta

forsteri were the fourth and fifth most abundant species, comprising just 4.1 and 1.7% of the

overall catch. In contrast, the percentage contributions of the five most abundant species caught

in the current study were relatively similar, with the Elongate Hardyhead, Sandy Sprat,

Yelloweye Mullet, Silverfish and the Spotted Hardyhead Craterocephalus mugiloides

contributing 19.1, 16.0, 12.7, 12.7 and 10.5%, respectively, to the overall fish catch.

Collectively, these species contributed 71% to the total catch.

Eight of the 11 most abundant species caught during 2008-10 ranked among the 11 most

abundant species in 1994 (Table 1). The three species that ranked in the top 11 in the present but

not earlier study included the Spotted Hardyhead and the Common Hardyhead Atherinomorous

vaigiensis, which are tropical species and were not caught during the earlier study. The third

species that ranked in the top 11 in the present, but not previous study was the Weeping Toadfish

Torquigener pleurogramma. This species was not superabundant, however, in either period,

26 ranking 12 in 1994 and 10 in 2008-10. The Western Australian Salmon Arripis truttaceus, which ranked 13 and was represented by 78 individuals in 2008-10, was not caught in 1994. In contrast to the above trends, the Southern Longfin Goby ranked first and contributed as much as 36.5% to the catches in the earlier period, compared with ranking sixth and contributing just 8.1% in the current period. The Southwestern Goby Afurcagobius suppositus ranked as high as sixth in 1994, but as low as 26 in 2008-10, when only three individuals were caught (Table 1). From the above trends, it is evident that the three most numerous species of gobies in the earlier period were far less abundant in the recent period.

Densities of commercial and recreational fish species

The mean densities (± 1 SE) of the most commercially and recreationally important fish species caught in the estuary in 1994 and 2008-10 were compared (Fig. 3). The mean densities of the Yelloweye Mullet increased significantly from 4.4 fish 100 m-2 in 1994 to 30.8 fish 100 m-2 in 2008-10. While not significant, the mean densities of King George Whiting also increased from 4.0 fish 100 m-2 in the earlier period to 9.4 fish 100 m-2 in the current study. The mean densities of the Yellow-fin Whiting and Sea Mullet were consistently low in both periods and exhibited no significant change between the two periods.

Numbers and percentage contributions of the different life cycle guilds

In terms of number of species, the percentage contributions of each life-cycle guild were relatively similar in the two periods (Table 2). There was, however, a slight increase in the contribution of marine estuarine-opportunist species between 1994 and 2008-10, due to the capture of the Western Australian Salmon, the Common Silverbiddy Gerres subfasciatus,

27 Sixspine Leatherjacket Meuschenia freycineti and the Serpent Eel Ophisurus serpens during

2008-10 but not 1994. In contrast, there were substantial differences in the numbers of individuals of species classified as estuarine & marine and estuarine residents between the two periods. The percentage contribution of species that can complete their life cycle in marine waters as well as estuaries (E&M) decreased from 53.6% in the earlier study to 26.8% in the current study, which was largely due to the decline in the abundance of the Southern Longfin

Goby. In contrast, the percentage contribution of those species that are restricted to estuarine waters increased from 7.8% in 1994 to 30.4% in 2008-10. This change was attributable mainly to an increase in the numbers of the Elongate Hardyhead and the colonisation of the Leschenault

Figure 3. Comparisons of the mean densities ± 1 SE of commercially and recreationally important species in nearshore waters of the Leschenault Estuary in 1994 (black circles) and 2008-10 (white circles).

28 Table 2. Comparisons of the numbers (N) and percentage contributions (%) of species and individuals of each life cycle guild caught in the nearshore waters of the lower and middle regions of the Leschenault Estuary in 1994 and 2008-10. Number of individuals represents totals following adjustment to a constant density, i.e. number of fish 100 m-2.

2008-10 1994 N % N % Species

Estuarine and marine 7 19.4 6 18.2 Estuarine residents 5 13.9 4 12.1 Marine estuarine-opportunists 19 55.6 16 51.5 Marine stragglers 5 11.1 7 18.2 Total 36 33

Individuals

Estuarine and marine 3276 26.8 6750 53.6 Estuarine residents 3723 30.4 984 7.8 Marine estuarine-opportunists 4984 40.7 4797 38.1 Marine stragglers 258 2.1 56 0.5 Total 12,241 12,587

Estuary by the Common Hardyhead and Spotted Hardyhead after 1994 and to effects of greatly reduced contributions by the Southern Longfin Goby, the major estuarine and marine species in the first period (Table 2). The percentage contribution of the numbers of individuals belonging to the marine straggler and marine-estuarine opportunist guilds were similar in the two periods.

Compositions of the fish faunas

A three-way PERMANOVA demonstrated that the compositions of the fish faunas in the middle and lower regions of the Leschenault Estuary in 1994, 2008/09 and 2009/10 differed significantly between seasons, years and regions and that there was a significant interaction between year and both season and region (Table 3). The COV was greater for season than for year, which, in turn, was far higher than that for region (Table 3). The COVs for the two-way interactions between year and season and between year and region were both greater than those

29 for region, but there was no significant interaction between either season and region or between year, season and region (Table 3).

Two-way crossed ANOSIM tests for year vs season and year vs region demonstrated that the compositions of the fish fauna in 1994, 2008/09 and 2009/10 were significantly different.

Pairwise comparisons across all seasons showed that the compositions in 1994 differed from those in both 2008/09 (R = 0.202, p = 0.1%) and 2009/10 (R = 0.387, p = 0.01%). Likewise, the composition in 1994 differed from that in both 2008/09 (R = 0.198, p = 0.01%) and 2009/10 (R =

0.257, p = 0.01%) across both regions. However, in both of the above tests, the fish compositions in 2008/09 and 2009/10 were not significantly different across either seasons (R = 0.022, p = 30.6%) or regions (R = 0.038, p = 12.0%).

Table 3. Mean squares (MS), pseudo F-ratios, components of variation (COV) and significance levels (p) for 3-way PERMANOVAs derived from the matrix constructed from the abundance of each species in samples collected seasonally in the lower and middle regions of the Leschenault Estuary in 1994, 2008/09 and 2009/10. df = degrees of freedom. Significant results are highlighted in bold.

Fish compositions Main Effects df MS Pseudo-F COV p Year 2 8392 3.327 207.3 < 0.001 Season 3 8576 3.400 286.9 < 0.001 Region 1 7953 3.153 128.9 < 0.001 Interactions Year × Season 6 4206 1.668 237.3 < 0.001 Year × Region 2 4956 1.965 171.9 0.010 Season × Region 3 4021 1.594 142.0 0.065 Year × Season × Region 6 2537 1.006 4.1 0.470 Residual 62 2522 2522.2

30 The ichthyofaunal compositions in all pairwise comparisons between each possible combination of seasons were significant, except for winter vs autumn (R = 0.072, p = 11.90%).

Furthermore, the comparisons between the compositions in each other pair of seasons yielded R- statistic values of between 0.186 and 0.358 (p = 0.01-0.20%), apart from in spring vs autumn, for which the R-statistic value was as high as 0.426 (p = 0.01%).

When the matrix, constructed from the mean abundances of the various fish species in the lower and middle regions of Leschenault Estuary in each season of each year, was subjected to

MDS ordination, all but one of the samples from 1994 lie immediately to the left of those from both 2008/09 and 2009/10 on the resultant ordination plot (Fig. 4a). In contrast, the samples from

2008/09 and 2009/10 intermingled on that plot. When the points were coded for season, and irrespective of year, all of the samples from spring formed a group to the left and/or above all of those for summer, while those for autumn and winter formed a widely-dispersed and discrete group that extended along the top and down the right side of the plot (Fig. 4b). There was no trend for the samples collected in 1994, 2008/09 and 2009/10 to group together according to region (Fig. 4c).

31

Figure 4. Non-metric multidimensional scaling (nMDS) ordination plot, constructed from a Bray-Curtis similartity matrix derived from the mean abundances of the various fish species in the lower and middle regions of the Leschenault Estuary in four successive seasons in 1994, 2008/09 and 2009/10. Data coded for a) year, b) season and c) region.

32 To explore visually the interaction between year and season, the matrix constructed from the mean abundances of each fish species in the Leschenault Estuary in each season of each year was subjected to nMDS ordination. On the ordination plot, the samples for 1994 formed a progression from summer downwards and to the left for autumn and winter and then upwards to spring (Fig. 5a). The samples for 2008/09 and 2009/10, which all lay to the right and/or below those for 1994, followed a similar clockwise progression on the ordination plot, with those for spring, summer and autumn of 2008/09 lying close to those of the corresponding seasons in

2009/10. The distance between the points for the samples in summer and autumn were far greater in 1994 than in either 2008/09 or 2009/10, which helps account for the significant interaction between year and season.

The interaction between year and region was explored visually by ordinating the matrix derived from the mean abundances of each species in each region in each year. On the resultant ordination plot, the samples from the lower regions in 1994, 2008/09 and 2009/10 lie above those for the middle region and in the same sequence according to year across the ordination plot

Figure 5. Non-metric multidimensional scaling (nMDS) ordination plot constructed from a Bray-Curtis similarity matrix derived from the mean abundances of the various fish species in a) each successive season and b) region of the Leschenault Estuary in 1994, 2008/09 and 2009/10. S = summer, A = autumn, W = winter, Sp = spring, L = lower region and M = middle region.

33 (Fig. 5b). The interaction is explained by the fact that, along the vertical axis, the sample from the middle region lies to the left of the lower region, whereas the reverse is the case in 2008/09 and 2009/10.

As the compositions of the fish faunas in 2008/09 and 2009/10 were not significantly different, the abundances of the various species in each region in those two years were pooled for comparisons with the composition in 1994. This pooling of data for the second period, i.e. 2008-

10, was also considered to be valid because the trends exhibited on the ordination plots by the samples for the four seasons (Fig. 5a) and the two regions (Fig. 5b) were similar.

The abundance data for the fish species in the two months in each season in 1994 were also pooled, thus combining the data for each of the seasons that were used for the above comparisons between years with those for the other additional data collected for those seasons

(see Materials and Methods). This created a far more robust data set for inter-period comparisons and produced identical numbers of “samples” for each period.

On the basis of a three-way PERMANOVA, the compositions of the fish faunas in the

Leschenault Estuary differed significantly between periods (1994 vs 2008-10), seasons and regions, with the COV for period and season being similar and well over twice that for region

(Table 4). There was a strong interaction between season and both period and region, with the

COV for these interactions being greater than that for region.

A two-way crossed ANOSIM test for period and season yielded Global R-statistic values of 0.441 and 0.341 for these factors, respectively. The trends exhibited by the various pairwise comparisons between seasons differed from those results for the same data for each of the three years separately in that the fish compositions were now significantly different between each pair

34 of seasons, with R-statistic values ranging from 0.219 to 0.576. The highest R-statistic value, and thus greatest temporal difference in fish compositions was that between spring and autumn.

Table 4. Mean squares (MS), pseudo F-ratios, components of variation (COV) and significance levels (p) for 3-way PERMANOVAs of the matrix constructed from the abundance of each species in samples collected from the lower and middle regions of the Leschenault Estuary, twice seasonally in 1994 and once seasonally in eight consecutive seasons in 2008-10. df = degrees of freedom. Significant results are highlighted in bold.

Fish Compositions Main Effects df MS Pseudo-F COV p Period 1 13686 6.943 427.1 0.0001 Season 3 7268 3.688 386.3 0.0001 Region 1 7005 3.554 183.5 0.0001

Interactions Period × Season 3 4636 2.352 388.6 0.0001 Period × Region 1 2196 1.114 16.4 0.3373 Season× Region 3 3416 1.733 210.7 0.0041 Period × Season × Region 3 2239 1.136 78.1 0.2536 Residual 40 1971 1971.1

A two-way crossed SIMPER for period and season demonstrated that the differences between the species compositions in 2008-10 and 1994 were attributable to consistently greater numbers of the Spotted Hardyhead (C. mugiloides), Weeping Toadfish (T. pleurogramma),

Yelloweye Mullet (A. foresteri) and Sandy Sprat (H. vittatus) and to consistently lower abundances of the Southern Longfin Goby (F. lateralis) and Prickly Toadfish (C. brevicaudus)

(Table 5). The influence of season is considered separately for each period (see below).

Subjection of the matrix, derived from seasonal samples collected in 1994, to one-way

SIMPER demonstrated that the composition of the fish fauna in each season of that year was typified by the Southern Longfin Goby (F. lateralis) and that the composition in spring was

35 always distinguished from that in each of the other seasons by consistently greater abundances of

Sandy Sprat (H. vittatus), Spotted Pipefish (S. argus) and Silverfish (L. presbyteroides) (Table 6).

Furthermore, the composition in winter was always distinguished from that in each of the other seasons by consistently greater abundances of the Prickly Toadfish (C. brevicaudus). In the case of 2008-10,

SIMPER demonstrated that, although F. lateralis was consistently caught and helped to typify the fish faunas in each season, it did not distinguish between the faunas in any pair of seasons (Table 7). The only species to exhibit a consistent trend to distinguish between seasons in 2008-10 were the tropical atherinids, the Spotted Hardyhead (C. mugiloides) and the Common Hardyhead (A. vaigiensis), which were always in greater abundance in autumn than in any other season (Table 7).

Table 5. Species that typified (shaded) and distinguished (unshaded) the fish faunas of the Leschenault Estuary in 1994 and 2008-10 as detected by two-way crossed SIMPER. The period in which each species was most abundant is given in superscript for each pairwise comparison.

1994 2008-10 F. lateralis C. brevicaudus 1994 S. punctata S. burrus L. presbytroides F. lateralis 94 F. lateralis C. mugiloides 08-10 T. pleurogramma T. pleurogramma 08-10 C. mugiloides 2008-10 C. brevicaudus 94 H. vittatus A. foresteri 08-10 L. presbytroides H. vittatus 08-10 A. elongata

36 Table 6. Species that typified (shaded) and distinguished (unshaded) the fish faunas of the Leschenault Estuary in 1994 as detected by a one-way SIMPER. The season in which each species was most abundant is given in superscript for each pairwise comparison.

Summer Autumn Winter Spring F. lateralis Summer S. burrusA F. lateralis L. presbyteroidesA Autumn S. punctataS T. pleurogrammaA F. lateralisA F. lateralisW F. lateralisW F. lateralis C. brevicaudusW C. brevicaudusW C. brevicaudus Winter S. punctataS S. burrusA L. presbytroidesA T. pleurogrammaA F. lateralisSp F. lateralisSp H. vittatusSp F. lateralis H. vittatusSp H. vittatusSp C. brevicaudusW L. presbyteroides L. presbyteroidesSp S. argusSp S. argusSp S. punctata Spring S. argusSp L. presbyteroidesSp L. presbyteroidesSp M. cephalusSp S. burrusSp S. burrusSp S. punctataS

Table 7. Species that typified (shaded) and distinguished (unshaded) the fish faunas of the Leschenault Estuary in 2008-10 as detected by a one-way SIMPER. The season in which each species was most abundant is given in superscript for each pairwise comparison.

Summer Autumn Winter Spring H. vittatus Summer F. lateralis A. elongata C. mugiloidesA A. vaigiensis A. vaigiensisA F. lateralis S Autumn H. vittatus T. pleurogramma L. presbyteroidesS A. elongata S. punctataS H. vittatusS C. mugiloidesA F. lateralis S. punctataS A. vaigiensisA T. pleurogramma Winter A. elongataS F. lateralisW T. pleurogrammaS S. argusSp C. mugiloidesA H. vittatusSp H. vittatus A. vaigiensisA S. argusSp F. lateralis Spring H. vittatusSp G. marmoratusSp T. pleurogramma S. argusSp C. mugiloidesW G. marmoratus G. marmoratusSp

37 On the ordination plot, derived from the Bray-Curtis resemblance matrix constructed from the mean abundances of the various fish species in each season in each period, the samples for each period formed essentially two discrete groups (Fig. 6a). While the samples for each period changed in a cyclical manner, the points for spring and summer were far more widely dispersed in 1994 than in 2008-10, which helps explain the interaction between period and season. The samples from the middle and lower estuary followed a similar cyclical pattern on the ordination plot derived from the Bray-Curtis resemblance matrix constructed from the mean abundances of the various fish species in each season in each region (Fig. 6b). While the samples for autumn in both regions were closely opposed, those for other seasons, and particularly those for winter, were more widely separated.

Figure 6. Non-metric multidimensional scaling (nMDS) ordination plot, constructed from a Bray-Curtis similarity matrix derived from the mean seasonal abundances of the various fish species in each a) period and b) region of the Leschenault Estuary in 1994 and 2008/10. S = summer, A = autumn, W = winter, Sp = spring.

38 Numbers of species and densities of fish

Three-way PERMANOVA demonstrated that the number of species caught in the nearshore waters of the lower and middle region of the Leschenault Estuary in 1994 and 2008-10 differed significantly between periods, seasons and regions and that there was a significant interaction between period and season for both of those variables (Table 7). The COV was similar for each of the main effects, which were all slightly greater than that for the interaction.

In the case of density, the seasonal and region main effects and the interaction between period and season were significant, with the COV being greatest for the interaction (Table 7).

The mean number of species in each sample was less in the lower region, i.e. 7.9, the middle region, i.e. 10.6 (Fig. 7a). While the mean number of species in 1994 was similar to those in 2008-10 in autumn, winter and spring, that for summer in 2008/10 was approximately twice that of summer in 1994 (Fig. 7b). As with number of species, the density in the lower region of the estuary was less than that in its middle region (Fig. 8a). The mean density of fish in both the autumn and winter of both periods were similar (Fig. 8b). In contrast, it was greater in summer in

2008-10 than in 1994, whereas the reverse was true in spring.

Table 7. Components of variation (COV) and significance levels for three-way PERMANOVAs of the mean number of species and density of fishes in the lower and middle regions of the Leschenault Estuary in two periods, i.e. 1994 and 2008-10. df = degrees of freedom. * p < 0.05, ** p < 0.01, *** p < 0.001.

Main effects Period (P) Season (S) Region (R) Residual df 1 3 1 94 Number of species 0.0955*** 0.1064*** 0.0970*** 0.1622 Density -0.0200 0.2791** 0.2438** 0.9466 Two- and three-way interactions P × S P × R S × R P × S × R df 3 1 3 3 Number of species 0.0795** -0.0089 0.0267 -0.0057 Density 0.5800** -0.0445 -0.0503 0.1400

39

Figure 7. Back-transformed mean numbers of species ± 95% confidence limits recorded in the nearshore waters of the Leschenault Estuary in a) each region during 1994 and in 2008-10 combined, and in b) each season during 1994 (black circles) and 2008-10 combined (white circles).

40

Figure 8. Back-transformed mean densities of fish ± 95% confidence limits recorded in the nearshore waters of the Leschenault Estuary in a) each region during 1994 and in 2008-2010 combined, and in b) each season during 1994 (black circles) and 2008-10 (white circles).

41 Sex ratio and minimum legal carapace width for Blue Swimmer Crabs

A total of 694 Blue Swimmer Crabs, Portunus pelagicus, was caught in eight seasons of sampling using 21.5 (182 crabs) and 41.5 m (318 crabs) long seine nets and in four seasons of sampling using crab pots (194 crabs). Two hundred and fourteen of these crabs were females,

470 were males and 10, all of which had a carapace width (CW) < 30 mm, were unable to be sexed macroscopically. The ratio of females to males exceeded parity in the samples collected by each of the three sampling methods (all p < 0.05). Thus, the female to male ratio of Blue

Swimmer Crabs caught using the 21.5 and 40.5m seine nets and crab pots were 1:1.4, 1:1.3 and

1:26.7, respectively, which were all significantly different (p < 0.05) from zero (χ2 = 5.4, 6.5 and

167.0, respectively). The overall ratio of females to males (1:2.2) was significantly different

(p < 0.001) from zero (χ2 = 95.8). Among the 694 Blue Swimmer Crab caught throughout the

Leschenault Estuary, the CW of only 59 were > 127 mm, the minimum legal carapace width for retention of P. pelagicus in Western Australia. Of the 59 Blue Swimmer Crabs with a CW > 127 mm, only five were female.

Densities and catch rates of Blue Swimmer Crabs

No Blue Swimmer Crabs were caught when using a 21.5 m seine net in any of the four regions of the Leschenault Estuary in the winters of either 2008 or 2009, when water temperatures were at their minima and salinities were relatively low (Fig. 9). Blue Swimmer

Crabs were only caught in the apex region in spring 2008 and, even then, their numbers were very low, which may reflect the extreme shallowness of the water in this region. Densities in each of the other three regions peaked in either spring or summer (Fig. 9). The above trends for the densities of Blue Swimmer Crab in the lower and middle regions of the estuary, derived from

42 data obtained using a 21.5 m seine net, were paralleled by those derived from data obtained using the 41.5 m seine net. Thus, no Blue Swimmer Crabs were caught with the 41.5 m seine net in the winters of either 2008 or 2009 and the densities likewise tended to peak in either spring or summer (Fig. 9). On the basis of data derived from both net types, the densities of Blue

Swimmer Crab in the lower and middle region were not significantly different (p > 0.05).

Blue Swimmer Crabs were caught in pots set in offshore, deeper waters in the lower and middle regions of the estuary in each of the four seasons of sampling except for the winter of

2009 in the lower region and the spring of 2009 in the middle region (Fig. 10). The mean catch rates in both regions were greatest in summer and autumn. They were also sometimes substantial, exceeding ~ 5 crabs pot-1day-1 in the summer of 2010 in the lower region and summer and autumn of 2010 in the middle region, with the mean catch rates on those three occasions being 7.6, 6.8 and 5.3 crabs pot-1day-1, respectively (Fig. 10).

Carapace width frequency distributions for Blue Swimmer Crabs

The CW of Blue Swimmer Crabs caught using the 21.5 and 41.5 m seine nets and crab pots ranged from 14 to 120 mm ( = 70.5 mm), 15 to 139 mm ( = 84.1 mm) and 98 to 153 mm

( = 120.8 mm), respectively. Thus, the three sampling methods collectively caught a wide size range of Blue Swimmer Crabs, including young juveniles and mature adults.

The carapace width frequency histograms for summer and autumn 2010 demonstrate that pots select larger Blue Swimmer Crabs than either the 21.5 or 41.5 m seine nets (Fig. 11). Thus, in those two seasons, the CW of the majority of Blue Swimmer Crabs taken in pots exceeded 100 mm, whereas those of only a few of the crabs taken in the seine nets were > 115 mm CW. This parallels the situation recorded in the Leschenault Estuary in 1996-98, when the

43

Figure 9. Mean seasonal salinities (black circles) and water temperatures (grey circles) ± 1SE and mean seasonal densities ± 1 SE of Blue Swimmer Crabs Portunus pelagicus caught using a 21.5 m and 41.4 m seine net in shallow, nearshore waters of the lower, middle, upper and apex regions of the Leschenault Estuary between winter 2008 and autumn 2010.

44

Figure 10. Mean seasonal salinites (black circles) and water temperatures (grey circles) ± 1 SE and mean seasonal catch rates ± 1 SE of Blue Swimmer Crabs Portunus pelagicus caught using crab pots in deeper, offshore waters of the lower and middle regions of the Leschenault Estuary between winter 2009 and autumn 2010.

45 CW of virtually all of the Blue Swimmer Crabs caught by pots exceeded 80 mm, while those of numerous crabs taken in seine nets in certain months, i.e. October to December, were less than this size (Potter and de Lestang, 2000). This gear selectivity thus needs to be borne in mind when interpreting the trends exhibited by the carapace width frequency data for sequential months, especially as pots were not used for sampling Blue Swimmer Crabs during the first four seasons of sampling. Yet, it is clear that the trends in 2009-10 are consistent with those exhibited by the corresponding frequency data in 1996-1998.

The study conducted in 1996-98, which involved sampling in Koombana Bay outside the estuary as well as in the estuary, demonstrated that the presumptive zoeae are released from the females in mid-spring to late summer (Potter and de Lestang, 2000). While a few small Blue

Swimmer Crabs, i.e. with CWs < 40 mm, entered the estuary in the ensuing months, substantial numbers of the 0+ age class did not start appearing in this system until spring.

During the present study, the late 0+ age class also appeared in numbers in the estuary in spring 2009 and at a similar size as in the previous study, i.e. in the CW range of 30-80 mm, and after having not been present in the winter (Fig. 11). The CWs of that cohort in the seine net samples increased rapidly to between mainly 65 and 105 mm in summer to 80 to 120 mm in autumn. Very few small early the 0+ recruit appeared in the summer and autumn of 2009

(Fig. 11). However, unlike the situation in 2008, the immigration of substantial numbers of late

0+crabs in 2009 was delayed from spring until summer, at which time they were of a similar size to those of the corresponding age class, i.e. early 1+, caught in the estuary in the previous summer. In contrast to the situation in autumn 2009, the seine net samples in autumn 2010 contained appreciable numbers of 0+ cohorts which, from their low numbers in the previous season, must have entered the estuary from Koombana Bay in the intervening period (Fig. 11).

46

Figure 11. Frequency histograms for the carapace widths of Blue Swimmer Crabs Portunus pelagicus caught in shallow, nearshore waters of the lower and middle regions of the Leschenault Estuary using 21.5 and 41.5 m seine nets between winter 2008 and autumn 2010 and from deeper, offshore waters using crab pots between winter 2009 and autumn 2010.

47 The second main mode in summer 2010, representing crabs with CWs ranging predominantly between 110 and 135 mm, is assumed to represent Blue Swimmer Crabs that are about two years old (Fig. 11). As this age class was poorly represented in the previous winter and spring, the members of this age class must have recently entered from outside the estuary. The presence of substantial numbers of 0+ crabs in autumn, representing crabs with CWs peaking at

55 to 70 mm, contrasts with the paucity of individuals of this age class in the previous year

(Fig. 11).

Historical comparisons of densities and catch rates of Blue Swimmer Crabs

Three-way PERMANOVA, derived from the matrix constructed from the seasonal densities of Blue Swimmer Crabs in the lower and middle regions in 1996/97, 1997/98, 2008/09 and 2009/10, demonstrated that the density of Blue Swimmer Crabs were significantly related to season and that there was a significant interaction between this variable and year, with the COV being greatest for season (Table 8). On the basis of data derived from seine netting in the two seasons when crab abundance was greatest (summer and autumn), the densities of Blue

Swimmer Crabs in 1996/97 and 2009/10 were appreciably greater than in 1997/98 and 2008/09

(Fig. 12).

Three-way PERMANOVA, derived from the matrix constructed from the seasonal catch rates of Blue Swimmer Crabs in the lower and middle regions in 1996/97, 1997/98 and 2009/10, demonstrated that the catch rates of Blue Swimmer Crabs were significantly related to year and season and that there was a significant interaction between these variables with the COV being greatest for season followed by year and then the interaction (Table 9). As with densities

(Fig. 12), the catch rates of Blue Swimmer Crabs in pots were also greater in 1996/97 and

48 2009/10 than in 1997/98 (Fig. 13). These data demonstrate that the strength of recruitment into the estuary varies among years and that the abundance in the most recent year of sampling was relatively high.

Table 8. Mean squares (MS), pseudo F-ratios, components of variation (COV) and significance levels (p) for 3-way PERMANOVAs of the matrix constructed from the densities of Blue Swimmer Crabs (100m-2) caught seasonally in nearshore waters of the lower and middle regions of the Leschenault Estuary, seasonally in 1996/97, 1997/98, 2008/09 and 2009/10. df = degrees of freedom. Significant results are highlighted in bold.

Main effects df MS Pseudo-F COV P Year 3 1.172 1.960 0.024 0.131 Season 3 7.675 12.838 0.295 0.000 Region 1 0.572 0.957 -0.001 0.339 Interactions Year × Season 9 1.696 2.836 0.183 0.007 Year × Region 3 0.710 1.188 0.009 0.327 Season × Region 3 0.712 1.191 0.010 0.325 Year × Seasons × Region 9 1.089 1.821 0.164 0.080 Residual 64 0.598 0.598

Table 9. Mean squares (MS), pseudo F-ratios, components of variation (COV) and significance levels (p) for 3-way PERMANOVAs of the matrix constructed from the catch rates of Blue Swimmer Crabs (crabs pot-1day-1) caught seasonally in offshore waters of the lower and middle regions of the Leschenault Estuary in 1996/97, 1997/98 and 2009/10. df = degrees of freedom. Significant results are highlighted in bold.

Main effects df MS Pseudo-F COV p Year 2 18.371 16.291 0.269 0.000 Season 3 42.266 37.479 0.857 0.000 Region 1 0.043 0.039 -0.011 0.847 Interactions Year × Season 6 6.307 5.593 0.324 0.000 Year × Region 2 2.096 1.859 0.030 0.156 Season × Region 3 1.337 1.186 0.009 0.321 Year × Seasons × Region 6 1.822 1.615 0.087 0.145 Residual 168 1.128 1.128

49

Figure 12. Seasonal mean densities ± SE of Blue Swimmer Crabs Portunus pelagicus caught using baited crab pots in nearshore, shallow waters of the lower and middle regions of the Leschenault Estuary between winter and autumn of 1996/97 (black circles), 1997/98 (grey circles), 2008/09 (white circles) and 2009/10 (dark grey circles).

Figure 13. Seasonal catch rates ± SE of Blue Swimmer Crabs Portunus pelagicus caught using baited crab pots in nearshore, shallow waters of the lower and middle regions of the Leschenault Estuary between winter and autumn of 1996/97 (black circles), 1997/98 (grey circles) and 2009/10 (white circles).

50 Discussion

This study demonstrated that, on the basis of seasonal sampling, the species compositions of the fish faunas of the Leschenault Estuary in two sequential 12 month intervals in the late

2000s, i.e. in 2008/09 and 2009/10, were not significantly different. This meant that the composition data for 2008/09 and 2009/10 could be pooled and considered representative of the late 2000s for comparisons with those derived from seasonal sampling in 1994, i.e. the mid-

1990s.

Although care must be exercised in drawing conclusions from the comparisons between the compositions in the late 2000s and a single year in the mid-1990s, it is clearly evident from such comparisons that the ichthyofaunal compositions in the two periods were similar in some respects. It was particularly notable, for example, that the Elongate Hardyhead, Sandy Sprat,

Yelloweye Mullet and Silverfish each ranked among the five most abundant species in both periods. It was also striking that the mean densities of fishes in the two periods were not significantly different in any of the four seasons and the same was true for the mean number of species in three of the four seasons.

The most conspicuous difference between the compositions in the two periods was the presence of substantial numbers of the Spotted Hardyhead and Common Hardyhead in samples from the late 2000s and their absence in those collected in the mid-1990s. It is thus clearly relevant that both of these are tropical species and that water temperatures along the coast have increased by 0.02 °C per annum over the 50 years up to the mid-2000s, with the increase being most marked between 1985 and 2004, i.e. 0.026 to 0.034°C per annum (Pearce and Feng, 2007). Thus, some individuals of these two species must presumably have been transported southwards from waters immediately to the north. In this context, it is relevant that

51 previous studies have shown that both of these species are at least moderately abundant in both the Swan-Canning Estuary (Loneragan et al., 1989) and Peel-Harvey Estuary (Potter et al.,

1983).

A second major difference between the compositions of the fish faunas in the mid-1990s and late 2000s resides in the substantial reductions that occurred in the densities of three species of gobies between these two periods. This was so pronounced in the case of the Southern

Longfin Goby and Southwestern Goby that its overall density declined by factors of 4.8 and 65 times, respectively, between the two periods. Furthermore, the ranking and contribution by abundance of the Southern Longfin Goby fell from first and 36.5 % in the mid-1990s to sixth and only 8.1% in the late 2000s. As the Southern Longfin Goby prefers sandy substrates (Gill and

Potter, 1993), and the contribution of fine particles to the sediment is now relatively high in the

Leschenault Estuary (Kilminster, 2010), the current benthic environment of this system could be leading to a greater tendency for the gills of this goby species to become clogged when it exhibits the shimmying behaviour that leads to it being largely covered and thus protected by (Gill and Potter, 1993). Such a sedimentary explanation cannot account, however, for the decline in the abundance of another goby species, the Southwestern Goby, as this species elsewhere is most abundant in areas where the substrate is silt/clay (Gill and Potter, 1993). Thus, the decline in the abundances of both species may reflect a change in some characteristic(s) of the benthos that affects goby species in general or that conditions in 1994 were particularly suitable for spawning and recruitment success in 1994. The latter possibility is consistent with the fact that the Southern Longfin Goby ranked only tenth and sixth in studies of the Swan-

Canning and Peel-Harvey estuaries (Loneragan et al., 1986; Loneragan et al., 1989; Loneragan and Potter, 1990), i.e. similar to the late 2000s in the Leschenault Estuary, and contributed only

52 1.9 and 3.8%, respectively, to the total catches in those systems. Furthermore, the percentage contribution rose only to 4.5 and 3.5%, when it was restricted to the preferred habitats of this species in the lower and middle regions of the Swan-Canning Estuary.

From a fish community perspective, it is pertinent that the Southern Longfin Goby and the Sandy Sprat were far less dominant in the late 2000s than the mid-1990s, as reflected by their collective contributions of 35 vs 69% to the total catches, respectively, and that the Spotted

Hardyhead and Common Hardyhead were moderately abundant in the late 2000s and were absent in the mid-1990s. These trends help account for the diversity and evenness of the fish fauna being greater in the late 2000s than the mid-1990s.

From the above, there is no clear evidence that the environment for the fish fauna in the

Leschenault Estuary has undergone deleterious changes. At the same time, it must be recognised that three important bottom-living species had declined in abundance and that this may reflect a deterioration in the benthic environment.

The trends exhibited by the densities and catch rates of Blue Swimmer Crabs with time of occurrence and space within the estuary in the late 2000s broadly parallel those recorded for this species in the 1990s (Potter and de Lestang, 2000). Thus, this species attainted their highest densities in the lower and middle estuary in summer and autumn, when salinities were at their maximum, and this species rarely caught in the apex of the estuary and only infrequently in the upper estuary.

The interannual trends exhibited by the densities of crabs were essentially the same as those of catch rates. Thus, the densities and catch rates in 1996/97 and 2009/10 were similar and greater than those recorded in 1997/98. These data thus demonstrate that the abundances of crabs

53 in the Leschenault Estuary vary conspicuously between years and provide no evidence that the crab assemblage has declined in abundance in recent years.

Acknowledgements

Our gratitude is expressed to friends and colleagues who helped with sampling. Funding was provided by the South West Development Commission, Western Australian Marine Science

Institution and Murdoch University.

References

Anderson, M. J., Gorley, R. N. and Clarke, K. R. (2008). PERMANOVA+ for PRIMER: Guide to software and statistical methods. Plymouth, UK: PRIMER-E.

Beckwith, J. (2008). Lower Collie River social values. Beckwith Environmental Planning Pty

Ltd, Western Australia, 35 pp.

Brearley, A., 2005. Ernest Hodgkin‟s Swanland: Estuaries and coastal lagoons of south-western

Australia. University of Western Australia Press, Crawley, Western Australia, 550 pp.

Clarke, K. R., (1993). Non-parametric multivariate analyses of changes in community structure.

Australian Journal of Ecology 18: 117–143.

54 Clarke, K. R. and Gorley, R.N. (2006). PRIMER v6: User manual/tutorial. PRIMER-E,

Plymouth, 190 pp.

Clarke, K. R. and Green, R. H. (1988). Statistical design and analysis for a 'biological effects' study. Marine Ecology Progress Series 46: 213-226.

Clarke, K. R. and Warwick, R. M. (2001). Change in marine communities: an approach to statistical analysis and interpretation, 2nd edition. Plymouth: PRIMER-E.

Clarke, K.R., Chapman, M. G., Somerfield, P. J. and Needham, H. R. (2006) Dispersion-based weighting of species counts in assemblages analyses. Marine Ecology Progress Series 320: 11-

27.

Elliott, M., Whitfield, A. K., Potter, I. C., Blaber, S. J., Cyrus, D. P., Nordlie, N. G. and

Harrison, T. D. (2007). The guild approach to categorizing estuarine fish assemblages: a global review. Fish and Fisheries 8: 241–268.

Gill, H. S. and Potter, I. C. (1993). Spatial segregation amongst goby species within an

Australian estuary, with a comparison of the diets and salinity tolerance of the two most abundant species. Marine Biology 117: 515-526.

55 Hillman, K., McComb, A. J., Bastyan, G. and Paling, E. (2000). Macrophyte abundance and distribution in Leschenault Inlet, an estuarine system in south-western Australia. Journal of the

Royal Society of Western Australia 83: 349-355.

Kelsey, P. and Hall, J. (2010). Nutrient loads, status and trends in the Leschenault Catchment,

Water Science technical series, Report no. 9, Department of Water, Western Australia.

Kilminster, K. L. (2010). Sediment quality in three south-western Australian estuaries, Water

Science technical series, Report no. 18, Department of Water, Western Australia.

Loneragan, N. R. and Potter, I. C. (1990) Factors influencing community structure and distribution of different life-cycle categories of fishes in shallow waters of a large Australian estuary. Marine Biology 106: 25-37.

Loneragan, N. R., Potter, I. C., Lenanton, R. C. J. and Caputi, N. (1986). Spatial and seasonal differences in the fish fauna in the shallows of a large Australian estuary. Marine Biology 92:

575-586.

Loneragan, N. R., Potter, I. C. and Lenanton, R. C. J. (1989). Influence of site, season and year on contributions made by marine, estuarine, diadromous and freshwater species to the fish fauna of a temperate Australian estuary. Marine Biology 103: 461-479.

56 McComb, A. J., Qiu, S., Paling, E. I. and Hill, N. A. (2000). Sediments of Leschenault Inlet: a comparison with other estuaries in south-western Australia. Journal of the Royal Society of

Western Australia 83: 275-284.

McKenna, M. (2000). The Leschenault Inlet/Estuary: a changing environment. Department of

Environment, Western Australia. 24 p.

McKenna, M. (2007). The Leschenault estuarine system, south-western Australia. Condition statement and recommendations for management, Department of Water, Western Australia.

Pearce, A. and Feng, M. (2007). Observations of warming on the Western Australian continental shelf. Marine and Freshwater Research 58: 914-920.

Potter, I. C. and Hyndes, G. A. (1999). Characteristics of the ichthyofaunas of southwestern

Australian estuaries, including comparisons with holartic estuaries and estuaries elsewhere in temperate Australia: a review. Australian Journal of Ecology 24: 395-421.

Potter, I. C. and de Lestang, S. (2000). Biology of the blue swimmer crab Portunus pelagicus in

Leschenault Estuary and Koombana Bay, south-western Australia. Journal of the Royal Society of Western Australia 83: 443-458.

57 Potter, I. C. Loneragan, N. R., Lenanton, R. C. J. Chrystal, P. J. and Grant, C. J. (1983).

Abundance, distribution and age structure of fish populations in a Western Australian estuary.

Journal of the Zoological Society of London 200: 21-50.

Potter, I. C., Chalmer, P. N., Tiivel,, D. J., Steckis, R. A., Platell, M. E. and Lenanton, R. C. J.

(2000). The fish fauna and finfish fishery of the Leschenault Estuary in south-western Australia.

Journal of the Royal Society of Western Australia 83: 481-501.

Raines, J., Youngson, K. and Unno, J. (2000). Use of the Leschenault Inlet estuary by waterbirds. Journal of the Royal Society of Western Australia 83: 503-512

Semeniuk, V. and Meagher, T. D. (1981). The geomorphology and surface processes of the

Australind-Leschenault Inlet coastal area. Journal of the Royal Society of Western Australia 64:

33-51.

Warwick, R. M. and Clarke, K. R. (1995). New „biodiversity‟ measures reveal a decrease in taxonomic distinctness with increasing stress. Marine Ecology Progress Series 129: 301-305.

58